Vehicle tracking

Abstract

The present invention relates to a method and system for accurately predicting future trajectories of observed objects in dense and ever-changing city environments. More particularly, the present invention relates to the use of prior trajectories extracted from mapping data to estimate the future movement of an observed object. As an example, an observed object may be a moving vehicle. Aspects and / or embodiments seek to provide a method and system for predicting future movements of a newly observed object, such as a vehicle, using motion prior data extracted from map data.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS[0001]The present application is a continuation of International Application No. PCT / GB2019 / 050514 filed Feb. 25, 2019, which claims priority to Great Britain Patent Application No. 1804195.4 filed Mar. 15, 2018 and Great Britain Patent Application No. 1810796.1 filed Jun. 29, 2018, which are hereby incorporated by reference herein.FIELD[0002]The present invention relates to a method and system for accurately predicting future trajectories of observed objects in dense and ever-changing city environments. More particularly, the present invention relates to the use of prior trajectories extracted from mapping data to estimate the future movement of an observed object. As an example, an observed object may be a moving vehicle.BACKGROUND[0003]A fundamental task of robotics perception and planning in dynamic environments is the ability to predict future evolution of the situation around the robot. For example, autonomous vehicles need to know about ...

AI Technical Summary

Benefits of technology

The patent describes a system that can predict the future position of an object, like a vehicle, based on its current position and information about other objects or vehicles in the area. This is useful for making accurate predictions without needing specific knowledge of the area's infrastructure or motion models. The system can use cheap visual sensors, like mobile phone cameras, to collect data on city-scale motion patterns and environmental information, which can be used to predict the future trajectories of newly observed vehicles. The system also uses information from structure from motion to improve the accuracy of future position predictions. The prediction system is modular, allowing for easy updates to the environment knowledge database. Overall, the system can make accurate predictions of object future position without needing specific knowledge of the surrounding area.

Problems solved by technology

Complex environments such as urban city traffic present significant challenges when it comes to such planning and perception.

However, a common disadvantage is that they often generalise the vast complexity of real world scenarios, such as busy intersections or turns, resulting in unreliable predictions.

Similarly, the motion of vehicles in complex scenarios cannot usually be predicted reliably using simple motion models like linear extrapolation, especially if the prediction horizon is greater than a few seconds.

However, the amount of work needed to produce such reliable maps and then to keep them updated is time consuming and heavily laborious.

Images